on 4/16/03 3:18 AM, Allen Briggs at briggs@wasabisystems.com wrote:

> On Wed, Apr 16, 2003 at 12:35:02AM +0200, Riccardo Mottola wrote:

>> Anyway the problems are threads, i tried posix and pthreads. in the gnustep
>> mailing list they pointed out to a thread problem confromance to posix in
>> both cases.
> 
> Hmmm...  Are you running a real recent -current?  If so, you should have
> access to our current native pthreads, which are aiming for POSIX, but
> which are also fairly new and still being fixed up.
I am using the official 1.6 kernel that went along with the distribution.
IIRC i recompiled it eliminating some bits like IDE or such. Nothing
-current for sure, if you think it is necessary to get something "current"
then explain me hoz to do it and I will try.
Howewer, the kernel I use has been proven quite stable except for the
infamous mmap() problem.

 
>> sendmsg.c: In function `__objc_get_forward_imp':
>> sendmsg.c:109: warning: control reaches end of non-void function
> 
> This doesn't look like a thread issue.
no it dosent, but honestly I'm too ignorant to really know what it fails.
That is only a warning, 9did jsut cut&paste the ending of the stdout) later
on it says:
thr-posix.c:101: warning: unused variable `params'
thr-posix.c: In function `__objc_thread_get_priority':
thr-posix.c:134: storage size of `params' isn't known
thr-posix.c:134: warning: unused variable `params'
thr-posix.c:140: warning: control reaches end of non-void function
gmake[1]: *** [shared_obj/m68k/netbsdelf1.6/gnu-gnu-gnu/thr-posix.o] Error 1
gmake: *** [libobjc.all.clibrary.variables] Error 2

which is in the thr-posix.c file, also here there are only warnings except
the fact that "storage size of 'parms' isn't known. Could that be the
problem?


> 
> I think you want POSIX and not PTHREADS, but we may be missing some of
> the sched priority functions.  Hmm..  They're there, but they're
> stubbed--they just return errors right now.

Yes, I'd prefer POSIX, they are the adviced way to compile gnustep.

thanks,
ric